Adaptive filtering for noise reduction in X-ray computed tomography

X-Ray Computed Tomography (CT) is one of the most important imaging modalities in radiology and has become an indispensable tool in many diagnostical and treatment assessment tasks. The radiation exposure is usually seen as the major drawback of the method. Of course, with respect to patients care, the use of lowest possible dose is desired. However, radiation exposure cannot be chosen arbitrarily low, due to its direct influence on image quality. With decreasing radiation dose, noise in the images increases, which makes a reliable diagnosis difficult or even impossible. In this thesis, two original approaches for noise reduction in reconstructed CT datasets are proposed and evaluated. The image quality improvement achieved by the new methods opens up a potential for dose reduction of 40-80 %, depending on the clinical task.